Flight Experiment of GBAS in Japan

Shinji Saitoh, Sonosuke Fukusima and Naoki Fujii

Electronic Navigation Research Institute,

Independent Administrative Institution, Japan

BIOGRAPHIES
Shinji Saitoh received his M. Eng. in Electronic
Engineering from the University of ElectroCommunications (UEC) in 1998. He joined Electronic
Navigation Research Institute (ENRI), Ministry of
Transport (MOT) in 1998. Currently, he is researching
in the area of GPS and GBAS at ENRI, independent
administrative institution (IAI).
Sonosuke Fukushima is a researcher at ENRI, IAI.
He received M. Eng. in Electronic Engineering from
the UEC. Since 1988 he is working in ENRI. He was
engaged in study on the development of MLS. His
research currently includes GBAS integrity for aircraft
approach and landing. He is a member of IEICE and
Japan Institute of Navigation.
Naoki Fujii is the manager of new landing system
section, ENRI, IAI. He received M. Sc. from Kobe
University in 1979. He was charged with development
of siting criteria of ILS, MLS and aircraft address
monitoring system. He is currently working in field
of development of GBAS.

ABSTRACT
Electronic Navigation Research Institute (ENRI) in
Japan have been developing and evaluating Ground
Based Augmentation System (GBAS) includes VHF
Data Broadcast (VDB) system. We conducted flight
experiments of the GBAS at Sendai airport. We
installed temporarily an experimental GBAS ground
station and executed the experiments in two types of

procedure. One is level flights including orbit and

arc flight patterns for checking the performance of the
stacked array antenna for null effects, and the other is
approach flights from 1,500 feet altitude for testing the
total performance of GBAS system.
The VDB system with horizontal polarization
generally has a null problem. Therefore we made a
stacked array antenna known for a solution of the null
problem and installed it at Sendai airport. The ground
station consisted of four sets of reference GPS antenna,
GPS receiver and personal computer (PC), and a
Work Station (WS) to generate GBAS augmentation
messages. The experiments were carried out with
Beechcraft B99 Airliner loaded a GBAS VDB receiver,
a GPS receiver and a PC for computing its real-time
differential GPS (DGPS) position and recording GBAS
data.
In the experiments, we also executed post-processed
kinematic carrier-phase differential positioning to get
reference positions. The results of flight to check
the null problem showed deep nulls were reduced
by the stacked array antenna and no message errors
were detected in GBAS service area.
In the
approach flights, we carried out real-time DGPS
positioning using GBAS augmentation information
(GBAS-DGPS) and computed difference between realtime GBAS-DGPS position fixes and post-processed
kinematic position fixes for evaluating positioning error
of GBAS. Final result shows total positioning accuracy
of our experimental system is nearly equal to CAT-III.
This paper introduces our GBAS equipments and the
flight test results about it.

Presented at the Institute of Navigation, ION-GPS-2001, SaltLake City, Sep.,2001

INTRODUCTION

receiving GBAS-VDB and a GPS receiving antenna

were mounted on a roof of the aircraft.

ENRI have been conducting studies, researches,

development of procedures and prototype systems,
test and evaluation on navigation systems so as to
provide Civil Aviation Bureau of the Ministry of
Land, Infrastructure and Transport, Japan (JCAB)
with technical materials for their planning and
implementation of navigation systems.

In the experiments, we executed two types of flight

procedure. One is a level flight including orbit and
arc flight patterns for checking the performance of the
stacked array antenna for null effects, and the other is
an approach flight from 1,500 feet altitude for testing
the total accuracy performance of the GBAS system.

We had developed and evaluated a Navigation

Augmentation Broadcast System (NABS) with Cband. But the C-band system had a problem of
shadowing and multipath by a body of the aircraft.
Now, ENRI have been developing and evaluating
GBAS VDB system. We has built up the preprototype model for testing system concept and carried
out flight trial for GBAS total system. We have
already developed VDB system and executed flight
test, and shown that the VDB system with horizontal
polarization has a null problem and the message errors
occur near null points in GBAS service area [1]. One
of solutions for the problem is adapting stacked array
antenna. Therefore we made a stacked array antenna
and installed it at Sendai airport in north-east region
of Japan , in January 2001.
We conducted flight experiments of the GBAS VDB
system with new VDB stacked antenna in 23 to 26
January and 23 July to 3 August 2001 at Sendai
airport. In the experiments, we installed temporarily
an experimental GBAS ground station. The ground
station has four sets of reference GPS antenna, GPS
receiver and PC, and a WS to generate GBAS
augmentation information. The experiments were
carried out with our experimental aircraft Beechcraft
B99 Airliner, shown in Figure 1. The B99 Airliner
loaded a GBAS VDB receiver, a GPS receiver and a
PC as a GBAS airborne station. A VHF antenna for

This paper describes overview of our experimental

GBAS-VDB system and results of flight experiments
for testing total performance of its system.

OVERVIEW OF ENRIS EXPERIMENTAL

SYSTEM
Figure 2 shows the configuration of the GBAS for
the flight experiment. It is a temporary system
for experiment and consists of GBAS ground station
included reference station and monitor station, and
GBAS airborne station. Its detail is described below.
In addition, stacked array VDB antenna is introduced.

Ground Subsystem
Figure 3 shows the photograph of the experimental
GBAS VDB system ground station main site. It was
installed in the Iwanuma branch of our institute next
to Sendai airport. It has five PCs for converting
from binary GPS receiver message of various type

LCD
PC
Ground station main site
(Iwanuma branch)
5 PCs

GPS ant.
GPS Rx
CDI
VDB Rx
Airborne Station

WS
VDB Tx
5 modems VDB ant.

4 GPS Rx
4 GPS ant.

Figure 1: ENRIs experimental Aircraft Beechcraft

B99 Airliner
A VHF antenna for receiving GBAS-VDB is
mounted on fore part of a roof. A GPS receiving
antenna is mounted on rear of the VHF antenna.

4 modems
Reference Station

GPS ant.
GPS Rx
modem
Monitor Station

Figure 2: Configuration of Experimental GBAS subsystem

Presented at the Institute of Navigation, ION-GPS-2001, SaltLake City, Sep.,2001

Figure 3: Experimental GBAS ground station main

site
From right: VDB transmitter, WS and PCs

with many manufactures to NMEA message, five

modems connecting with reference station and monitor
station and VDB transmitter connected to a new VDB
transmit array antenna, shown in Figure 7. The
reference station has four sets of GPS receiver (NovAtel
MiLLen. STD. ) , GPS receiving antenna (NovAtel
GPS600), and modem. The monitor station also has a
GPS receiver, GPS antenna and modem. A elevation
mask of all GPS receivers were set to 5 degrees through
the experiments.
The WS generate the GBAS augmentation messages
from pseudoranges, carrier-phase measurements and
ephemeris data obtained by four GPS receivers at
reference station. Moreover the WS checked the
computed position of monitor station using GPS data
from monitor station. The messages made by the
WS were broadcasted by the VDB system. In the
experiments, the system broadcasted GBAS type 1
messages at the rate of 2 Hz, and type 2, 4 and 5
messages at the rate of 0.1 Hz.
Figure 4 shows photographs of four GPS antennas for
GBAS reference station. All antennas were NovAtel
GPS600 antennas and settled in Sendai airport at
green area in front of Iwanuma branch of ENRI, along
A runway and A-1 taxi way. The height of each GPS
antennas were different, 1.7m to 4.0m, for reducing
multipath effect from ground, and separation of each
antennas were about 80m, formed a square. These
dispositions are shown in Figure 5.
Trimble 4000SSi and 5700 GPS receivers were installed
at ground station for calculating reference position
by post-processed kinematic carrier-phase differential
position fixes. A GPS antenna for these receivers was
mounted on top of VDB antenna tower.
The summary of GBAS reference station in the

Figure 4: Reference GPS antennas

Each height of antenna was 1.7m (upper left),
2.0m (upper right), 3.0m (lower left), 4.0m (lower
right).

Figure 5: Configuration of GBAS sites in Sendai

airport

Table 1: Summary of GBAS reference station

GBAS reference station
(GPS600L1/L2)
number of GPS antennas 4
disposition square
height 1.7m, 2.0m, 3.0m,
4.0m
separation about 80m
GPS receiver
NovAtel MiLLen. STD.
Elevation mask angle
5 degrees
kinematic reference station
GPS receiver
Trimble 4000SSi
Trimble 5700
antenna position
top of VDB tower

experiments was shown in Table 1.

The monitor station was installed away about 1km
from the reference station and it has a NovAtel
MiLLen. STD. GPS receiver, NovAtel GPS600 GPS
antenna and modem connected to the ground station
main site.
Figure 6 shows the pr gnd value in pseudorange
measurements of GBAS site in Sendai airport. The

Presented at the Institute of Navigation, ION-GPS-2001, SaltLake City, Sep.,2001

A4
B4

pr-gnd [m]

0.8

pr-gnd

0.6
0.4
0.2
0
0

10

20

30
40
50
60
Elevation angle [deg]

70

80

90

Figure 6: pr gnd of Sendai airport GBAS site and

GBAS designator values
black dashed curve is A4 (accuracy designator is A
with four ground receivers), the blue dashed curve
is B4 (accuracy designator is B with four ground
receivers) and the red solid curve is the pr gnd value
obtained at GBAS site in Sendai airport. We used
B4 value for computing GBAS position fixes and
protection level in the experiments.

Stacked array VDB antenna

The new VDB stacked array antenna, shown in Figure
7, consists of triple omni-directional antennas (similar
to VOR side-band Alford-loop antenna). Each height
of loop antennas is 15.5 meters (top), 10.5 meters
(middle) and 5.5 meters (bottom). And phase delays
were /2 radians (middle) and radians (bottom)
from phase of top antenna. The transmitting power
was 15 watts through the flight experiments, but
its capability of maximum transmitting power was
50 watts, because of reducing the interference on
television sets. The transmitting frequency was 108.5
MHz and modulation was D8PSK. Table 2 shows
summary of the new VDB stacked array antenna.

Table 2: Characteristics of VDB system with stacked

array antenna
frequency
108.5 MHz
transmit power
15W (max 50W)
Alford-loop
transmitting antenna
height from ground
power splitter ratio
phase delay [deg]
polarization

(similar to VOR side-band

ant.) ,

3 stacked array
5.5m , 10.5m , 15.5m
1
:
1
:
1
180 : 90 :
0
horizontal

Figure 7: New VDB stacked array antenna

Height of segment was 15.5m (top), 10.5m
(middle), 5.5m (bottom) each.
Another GPS antenna for kinematic reference was
mounted on top of this tower, height of 17.2m.

Airborne Subsystem
Figure 8 shows airborne station of our experimental
GBAS VDB system. The airborne station consists
of a GBAS-VDB receiver, NovAtel MiLLen. STD.
GPS receiver and PC for processing and recording the
GBAS and GPS data. The PC computes the aircraft
position using pseudoranges from GPS receiver and

Presented at the Institute of Navigation, ION-GPS-2001, SaltLake City, Sep.,2001

FLIGHT EXPERIMENTS

We conducted GBAS flight experiments in January

and July 2001. We executed two types of flight
procedure in the experiments. One is a level flight
including orbit and arc flight patterns for checking
the performance of the stacked array antenna for null
effects, and the other is approach flights from 1,500 feet
altitude for testing the total performance of the GBAS
system. We executed real-time DGPS positioning
using GBAS augmentation messages through the
approach flights.

Flight Experiment for VDB Stacked Array

Antenna
First, we conducted level flight of 8,000 feet altitude.
This flight includes two round trips from Sendai airport
to 55 nautical miles (NM) away to the direction of 110
clockwise from magnetic north. Figure 9 shows results
of this experiment.
A single VDB antenna with horizontal polarization
has null problem and caused message errors near null
points. Figure 10 shows result of previous experiment
with single Alford-loop VDB antenna on level flight of
8,500 feet altitude in June 1999.

GBAS pseudorange corrections from VDB receiver.

The positioning rate was 4Hz. Trimble receivers
(4000SSi and 5700) for computing carrier-phase DGPS
reference position were also installed onboard. Table 3
summarizes equipments loaded onboard in the aircraft.

Track
Message error
Input Level

8000
6000
4000
2000
0

10

Altitude [ft]

loaded onboard in B99

45

Figure 9: Results of level flight with 3 stacked array

antenna
in January 2001
The cadet blue area indicates the SARPs
requirement coverage of GBAS-VDB
10000

Table 3: Equipment
equipment
GBAS-VDB receiver
Personal Computer
GPS receiver
GBAS station
kinematic reference

15
20
25
30
35
40
Distance from VDB antenna [NM]

Track
Message Error
Input Level

8000
6000
4000
2000

for GBAS processing

NovAtel MiLLen.STD.
Trimble 4000SSi
Trimble 5700

-60
-70
-80
-90
-100
-110
-120
50

10

15
20
25
30
35
40
Distance from VDB antenna [NM]

45

-60
-70
-80
-90
-100
-110
-120
50

Receiver Input Level [dBm]

Figure 8: Experimental GBAS airborne subsystem

top: LCD display
middle: PC for GBAS processing
bottom: GBAS-VDB receiver

Altitude [ft]

10000

Receiver Input Level [dBm]

Figure 9 and Figure 10 show the receiving power level,

Figure 10: Results of level flight in previous experiment with Single Alford-loop antenna
in June 1999

Presented at the Institute of Navigation, ION-GPS-2001, SaltLake City, Sep.,2001

1000

-80

500

-100

0
-1

-120
0

Distance from R/W threshold [NM]

Figure 11: Results of approach flight with 3 stacked

array antenna
in January 2001
the altitude of aircraft, the positions of message error
occurred, and SARPs requirement coverage of GBASVDB. The red solid curve is the receiving power level,
the blue dashed line is the altitude of aircraft, the
black triangle is the position of error occurred and the
cadet blue area is SARPs requirement coverage. The
horizontal axis indicates distance from VDB antenna,
the left vertical axis displays the altitude of aircraft
and the right vertical axis displays the receiving power
level.
Figure 9 displays that there are no nulls and
no message errors occurred in SARPs requirement
coverage, while the null appears due to single Alfordloop antenna shown in Figure 10. Figure 9 shows
the lowest null caused in inbound flight with elevation
angle about 12.5 at the 6 NM distance from VDB
antenna.
We conducted approach flight from 1,500 feet altitude
for testing new VDB antenna performance at approach
phase of flight. Figure 11 shows an example. The red
solid curve is the receiving power level and blue dashed
curve is altitude track of aircraft. The horizontal axis
displays distance from the runway threshold and the
left vertical axis displays the altitude of aircraft and the
right vertical axis displays the receiving power level.
The receiving power has enough level to satisfy the
SARPs requirement and no errors were detected at the
SARPs requirement coverage.

Altitude [m]

-60

1.6
1.4
1.2
1
0.8
0.6

Track

-940

-920
-900
-880
Distance from R/W threshold [m]

-860

Figure 12: Post-processed kinematic positions on the

runway
mean=1.106 [m], =0.061 [m], MAXMIN=0.262
[m], 42 samples at 900 meters distance from
runway threshold.
aircraft trajectory and runway surface including height
of GPS antenna on aircraft, was within about a foot.
Figure 12 shows post-processed kinematic positions,
when aircraft was on the runway. The red solid line
is track of aircraft. The horizontal axis is distance
from the runway threshold and the vertical axis is
altitude from the runway threshold. At the 900 meters
distance point from runway threshold, we estimate the
positioning accuracy of post-processed kinematic was
6 centimeters (at 1 ) in the experiments.

Flight Experiment of Real-time GBAS-DGPS

positioning
We conducted approach flight for measuring GBAS
positioning accuracy in January and July 2001.
Figure 13 and Figure 14 show tracks of twenty-seven
times approach flights (touch-and-go flights) to B27
runway in July 2001. The horizontal axes of both figure

Lateral [m]

Track
Input Power Level

2000
1500

Input Power Level [dBm]

Altitude [ft]

2500

400
200
0
-200
-400
-1

Reference Position

1
2
3
4
5
6
Distance from R/W threshold [NM]

Altitude [ft]

Figure 13: Horizontal Aircraft track

In the experiments, we used post-processed kinematic
carrier-phase differential position fixes as reference
positions of aircraft.
Through flight in January
2001, we found bias errors in vertical direction and
found that post-processed kinematics carrier-phase
differential positioning jumped often, perhaps due to
cycle slip, when the attitude of aircraft changed. Then
we executed touch-and-go flight in July 2001, and we
selected available data by altitude of aircraft trajectory
on runway, when a separation, between altitude of

2500
2000
1500
1000
500
0
-1

Presented at the Institute of Navigation, ION-GPS-2001, SaltLake City, Sep.,2001

1
2
3
4
5
6
Distance from R/W threshold [NM]

Figure 14: Vertical Aircraft track

VPL95%
Z
Num. of SAT.

2
1
0
-1
-2

12
8
4
-1

1
2
3
4
5
Distance from R/W threshold [NM]

12
8
4

12
8
4

VPL95%
Z
Num. of SAT.

2
1
0
-1
-2
-1

Num. of Common GPS

1
2
3
4
5
Distance from R/W threshold [NM]

Num. of Common GPS

VPL95%, Z [m]
VPL95%, Z [m]

Num. of Common GPS

Track
Runway surface + Ant height
Z

12
8
4

Num. of Common GPS

2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1

VPL95%
Z
Num. of SAT.

2
1
0
-1
-2
-1

Distance from R/W threshold [NM]

-1

Distance from R/W threshold [NM]

VPL95%, Z [m]

Altitude, Z [m]

Figure 15 shows vertical error, track of aircraft and

B27 runway surface level. The horizontal axis shows
horizontal distance from the runway threshold and the
vertical axis shows vertical errors in meters and aircraft
altitude in meters. The red solid line displays vertical
errors of GBAS-DGPS, the blue dashed curve displays
altitude of aircraft track and the black solid curve
displays altitude of runway surface including height
of GPS antenna on aircraft. We selected kinematics
positioning data as references, when a separation of
blue and black curve was within about a foot at
touchdown point.

VPL95%, Z [m]

show distance from the runway threshold. The vertical

axis of the Figure 13 shows lateral shift from extended
runway center-line, while the vertical axis of the Figure
14 shows aircraft altitude. These data were selected by
aforementioned method.

VPL95%
Z
Num. of SAT.

2
1
0
-1
-2
-1

Distance from R/W threshold [NM]

Figure 15: Vertical error and aircraft track

The vertical error by real-time processing data, shown
in Figure 15, has a trend in accordance with distance
from runway threshold. Then we checked GBAS
parameters, and we found the tropospheric correction
parameters NR , in type2 message, was wrong. Figure
16 shows results of re-calculated GBAS-DGPS. The
axes are same in Figure 15 and the red solid line shows
vertical errors. No trend would be found in this figure.

Figure 17: Examples of vertical error and VPL95%

1st to 4th approaches in July 31 2001
Figure 17 and Figure 18 show examples of vertical
error and VPL95% . The VPL95% value indicates 95%
probability value of VPL and to be compared with
VAL95% . We bring in this value because we didnt have
enough samples for statistics analyzing. A definition
of VPL95% is below.

Z [m]

VPL95% =
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1

-1

1
2
3
4
5
Distance from R/W threshold [NM]

Figure 16: Corrected vertical error of re-calculated

GBAS-DGPS

2
VPLH0
Kf f md

The Kf f md is 5.847 (four references) for our

experiments. The horizontal axes are distance from
the runway threshold, the left vertical axes are VPL95%
and Z (vertical error) in meters and right vertical
axes are number of common GPS satellite. The green
dashed line is VPL95% , the red solid line is vertical
error and the blue dashed line is number of satellite
using GBAS positioning.
Through the experiments, we found the vertical errors
were always within VPL95% . When the number
of common satellite was changed, the vertical error

Presented at the Institute of Navigation, ION-GPS-2001, SaltLake City, Sep.,2001

Num. of Common GPS

12
8
4

Num. of Common GPS

12
8
4

Num. of Common GPS

12
8
4

Num. of Common GPS

VPL95%, Z [m]
VPL95%, Z [m]

-1

1
2
3
4
5
Distance from R/W threshold [NM]

VPL95%
Z
Num. of SAT.

2
1
0
-1
-2
-1

VPL95%, Z [m]

12
8
4

VPL95%
Z
Num. of SAT.

2
1
0
-1
-2

1
2
3
4
5
Distance from R/W threshold [NM]

VPL95%
Z
Num. of SAT.

2
1
0
-1
-2
-1

VPL95%, Z [m]

Distance from R/W threshold [NM]

VPL95%
Z
Num. of SAT.

2
1
0
-1
-2
-1

Distance from R/W threshold [NM]

Figure 18: Examples of vertical error and VPL95%

5th to 8th approaches in July 31 2001

experiences a step bias shift, shown in the 5th and the

8th approaches in Figure 18. But this shift was small,
less than few ten centimeters. Also the lateral errors
were within LPL95% , not shown in this paper.

Table 5: Positioning errors at monitor station

X
Y
Z

[m]
0.004
0.046
0.097

[m]
0.100
0.123
0.280
|| + 2 [m]
0.204
0.292
0.657

regard to carrier-phase differential position fixes as

reference of the approach flights. X indicates error
of along runway, Y indicates error of across runway
(lateral) and Z indicates error of vertical (altitude).
The value of RNP95% is required navigation system
error (95%) defined in SARPs [2] and AL95% value
is modified alert-limit value with 95% probability
value to be compared with 2 error of experimental
observations, computed with Kf f md multiplier defined
in SARPs [2].
We also analyzed data of the monitor station. The
results of the analysis show || + 2 = 0.657 [m] in
the vertical errors, shown in Table 5. This result was
better than at the airborne station , || + 2 = 0.691
[m] shown in Table 4. Its value was appropriate with
considerations of moving of aircraft and accuracy of
kinematic positioning.
Table 6 shows specifications of NSE 95%. The results
of the FAA flight test reported in the ION meeting
[4] show the vertical errors (|| + 2) was 1.6 to
2.3 meters at 100 feet above touchdown (CAT-II).
Their system positioning accuracy was nearly CATII accuracy and LAAS CAT-III. Table 7 shows vertical
and lateral navigation system error (NSE) at variant
distance from the runway threshold in the approach
of twenty-seven times touch-and-go flights to B27
runway with our experimental system. A distance of
900 meters correspond to CAT-I decision height, 300
meters correspond to CAT-II and 0 meters correspond
to CAT-III.

Table 4 shows results of GBAS-DGPS errors with

Table 4: Results of GBAS-DGPS Positioning Errors

(B27 R/W T&G, July 2001)
X
Y
Z
mean value [m] 0.207
0.012
0.009
std. deviation [m]
0.137
0.161
0.341
2 [m]
0.273
0.323
0.682
|| + 2 [m]
0.480
0.335
0.691
maximum value [m] 0.485
0.737 0.977
RNP95%
[m]

16.000
4.000
AL95%
[m]

13.682
3.421
number of samples
11623
11623
11623
approaches
27
27
27

Table 6: Specifications of NSE 95%

specification ac- specification according to GNSS cording to LAAS
draft SARPs [3]
MASPS
CAT I
200 feet decision height
16.0 m
16.0 m
Lateral
4.0 6.0 m
4.0 m
Vertical
CAT II
100 feet decision height
6.5 m
6.9 m
Lateral
1.7 m
2.0 m
Vertical
CAT III
50 feet decision height
3.9 m
6.2 m
Lateral
0.8 m
2.0 m
Vertical

Presented at the Institute of Navigation, ION-GPS-2001, SaltLake City, Sep.,2001

Table 7: NSE results, variant distance from R/W

threshold
Distance of 300m corresponds to 100ft height
above touchdown and distance of 900m corresponds to 200ft in case of gride path angle is 3 .
distance
0m
300m
900m
from R/W
threshold
CAT-III
CAT-II
CAT-I
vertical
[m]
0.001
0.002
0.002
[m]
0.389
0.397
0.359
|| + 2 [m]
0.778
0.795
0.721
lateral
[m]
0.040
0.038
0.026
[m]
0.175
0.166
0.166
|| + 2 [m]
0.390
0.371
0.358
number of
27
27
27
samples

of NEC Corp. for technical support and the personnel

who helped our experiments in Sendai airport.

REFERENCES
[1] S. Saitoh, S. Fukushima, N. Fujii; Results
of GBAS flight test (described in Japanese
language), Proceedings of the 32nd ENRI lecture
meeting, June, 2000.
[2] ICAO, SARPs, Annex 10, Nov., 2001. (publishing
schedule)
[3] Report of ICAO GNSSP WG/B, Appendix E,
Wellington New Zealand, Mar., 1998.
[4] John Warburton, Cliff Mackin, Ruben Velez;
Flight Test Results of the FAA Local Area Augmentation System Test Prototype, Proceedings of
the ION 53rd Annual Meeting , 1997.

The results of 95% probability vertical error (|| +

2) was 0.80 meters at 300 meters distance from
threshold and 0.79 meters above threshold, and CATII specification of vertical NSE is 1.7 meters and
CAT-III is 0.8 meters. Comparing these values, our
experimental system has enough accuracy using CATII and it has nearly CAT-III accuracy.

CONCLUSIONS
This paper presents the information of current GBAS
developing and evaluating status in Japan.
We clarified following two matters in this paper.
Using the stacked array VDB antenna is one of
the best solutions for the null problem according
to level flight experiments.
Positioning accuracy of our experimental GBASVDB system satisfies GBAS NSE specifications
for the CAT-II and the system accuracy is nearly
equal to CAT-III.
We will upgrade of reference GPS antennas and
check for integrity of our GBAS equipment by more
experiments.

ACKNOWLEDGMENTS
The authors would like to thank Mr. HIRAMA
Hideaki of Iwanuma branch, Mr. ISHIKAWA Kazunori
Presented at the Institute of Navigation, ION-GPS-2001, SaltLake City, Sep.,2001